Neutrality and the Evolvability of Boolean Function Landscape
EuroGP '01 Proceedings of the 4th European Conference on Genetic Programming
Finding needles in haystacks is harder with neutrality
Genetic Programming and Evolvable Machines
A comparison of bloat control methods for genetic programming
Evolutionary Computation
Representation and structural biases in CGP
CEC'09 Proceedings of the Eleventh conference on Congress on Evolutionary Computation
What makes a problem GP-hard? validating a hypothesis of structural causes
GECCO'03 Proceedings of the 2003 international conference on Genetic and evolutionary computation: PartII
Redundancy and computational efficiency in Cartesian genetic programming
IEEE Transactions on Evolutionary Computation
Reducing wasted evaluations in cartesian genetic programming
EuroGP'13 Proceedings of the 16th European conference on Genetic Programming
| Hi-index | 0.00 |
In this paper we examine how Cartesian Genetic Programming's (CGP's) method for encoding directed acyclic graphs (DAGs) and its mutation operator bias the effective length of individuals as well as the distribution of inactive nodes in the genome. We investigate these biases experimentally using two CGP variants as comparisons: Reorder, a method for shuffling node ordering without effecting individual evaluation, and DAG, a method for removing the concept of node position. Experiments were performed on four problems tailored to highlight potential search limitations, with further testing on the 3-bit multiplier problem. Unlike previous work, our experiments show that CGP has an innate parsimony pressure that makes it very difficult to evolve individuals with a high percentage of active nodes. This bias is particularly prevalent as the length of an individual increases. Furthermore, these problems are compounded by CGP's positional biases which can make some problems effectively unsolvable. Both Reorder and DAG appear to avoid these problems and outperform Normal CGP on preliminary benchmark testing. Finally, these new techniques require more reasonable genome sizes than those suggested in current CGP, with some evidence that solutions are also more terse.